Ahmet Aladag

An AI assistant for performing lead conversion using a chat box embedded into any webpage. I designed and developed the AI system and integrated it with the back and front end so our customers can embed the chat box into their system easily.

A web-based app that allows Amazon Business users to import their office addresses in bulk.

I was the owner of the product, and I gathered business requirements, prepared BRD, designed the software architecture, performed security analysis, and wrote extensive security documents, including a threat model. The project was carried out while working at Amazon.

A machine learning-based R&D project to predict whether the author of a post has expressed suicidal ideation in the post. The ultimate goal was to detect people at risk and provide them assistance. It was the proof of concept, the very first study and project on the topic. Our paper attracted great attention and was cited by over 50 other studies.

I used text-mining tools to extract features like sentiment analysis, LIWC (linguistic inquiry and word count), and bigrams. Then I used logistic regression, random forests, and support vector machine algorithms to build the models and make predictions.

Developed a major feature project for Twist. It allows users to import their Slack and Stride conversations and team members into Twist.

I was the back-end developer for the project, designed the architecture, and implemented the code in Python.

A command line-based app that allows you to manage your portfolio and perform trading. It enables you to display the average costs of coins based on your trade history. It also allows you to execute trades and manage multiple crypto exchanges.

A data science project that analyzes the public transportation (bus) network (consisting of routes and stops). The project highlights the potential bottlenecks in the traffic depending on the frequency of use.

I scraped the route and stop data from the Istanbul Municipality City SDK using Scrapy. Then I constructed, visualized, and analyzed the Istanbul public transportation network with R.

A blog project where I wrote about 70 blog posts in English for an Italian scraping company. The posts were about what scraping is, how it can be utilized in different industries, use cases, and how the data can be stored, processed, and analyzed.

A type of protein-protein interaction visualization (desktop) software. I developed the app portion for my master's thesis "Reliability Oriented Bioinformatic Network Visualization" tool.

It performs clustering on protein-protein interaction networks according to co-expression and co-ontology and then visualizes these graphs. Then it takes reliability scores into account during the visualization.

A research project to build an algorithm that would align protein interaction networks. I was involved in algorithm development and paper writing. Our paper was cited more than 130 times.

Given the protein-protein interaction (PPI) networks of a pair of species, a pairwise global alignment corresponds to a 1-to-1 mapping between their proteins. Based on the presupposition that such a mapping provides pairs of functionally orthologous proteins accurately, the results of the alignment may then be used in comparative systems biology problems such as function prediction/verification or construction of evolutionary relationships.

We provided a polynomial-time heuristic algorithm, SPINAL to solve this problem. We showed that based on the accuracy measures used in relevant work, our method outperforms the state-of-the-art algorithms.

Our algorithm does not suffer from scalability issues, as such accurate results are achieved in reasonable running times as compared with the benchmark algorithms.

A static site server that was written with the Crystal language and Kemal Web Framework. It compiles the Markdown content in your Dropbox folder into HTML with Pandoc every X minutes.

Optionally, it watches for changes in your Dropbox folder, then compiles all your Markdown files into HTML with Pandoc when any change is detected. It also provides a hook URL so that it can copy and rebuild the Markdown files upon a change in your Dropbox folder.

Access can be restricted with passwords and permissions. Each folder can be assigned special permissions via files in the config folder.

A Java library for ranking nodes in a graph using Apache Giraph. I implemented LinkRank, HostRank, and TrustRank algorithms to work on a distributed cluster.

We ran my algorithms on a cluster of 100+ nodes running for a search engine. The rankings were used to sort the search results.

A Python library implementing unidirectional and bidirectional friendships on Neo4J.

I Implemented the whole project. It allows you to construct a social network and provide a faster experience for querying friendship relationships. It supports friendship requests and approvals as well.

A text mining project that analyzes terms searched on an English-Turkish and Turkish-English dictionary to construct a semantic network based on co-occurrence.

I implemented the code and ran it on the data provided by the leading dictionary Zargan.

With the assumption of the fact "terms searched consecutively are more likely to be related," I created a co-occurrence network and with some pruning, the network turned out to be a semantic network suggesting relationships between the nodes. The network is suitable for a word representation approach in machine learning projects.

A web-based app for recruiting employees in a particular industry. I am the lead engineer and the back-end developer developing REST APIs for the React front end to consume. I designed the architecture and implemented the back end.